1 // Copyright 2014 The Chromium Authors
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
5 // This implementation doesn't use ICU. The ICU macros are oriented towards
6 // character-at-a-time processing, whereas byte-at-a-time processing is easier
7 // with streaming input.
9 #include "base/i18n/streaming_utf8_validator.h"
11 #include "base/check_op.h"
12 #include "base/i18n/utf8_validator_tables.h"
17 uint8_t StateTableLookup(uint8_t offset) {
18 DCHECK_LT(offset, internal::kUtf8ValidatorTablesSize);
19 return internal::kUtf8ValidatorTables[offset];
24 StreamingUtf8Validator::State StreamingUtf8Validator::AddBytes(const char* data,
26 // Copy |state_| into a local variable so that the compiler doesn't have to be
27 // careful of aliasing.
28 uint8_t state = state_;
29 for (const char* p = data; p != data + size; ++p) {
30 if ((*p & 0x80) == 0) {
33 state = internal::I18N_UTF8_VALIDATOR_INVALID_INDEX;
36 const uint8_t shift_amount = StateTableLookup(state);
37 const uint8_t shifted_char = (*p & 0x7F) >> shift_amount;
38 state = StateTableLookup(state + shifted_char + 1);
39 // State may be INVALID here, but this code is optimised for the case of
40 // valid UTF-8 and it is more efficient (by about 2%) to not attempt an
41 // early loop exit unless we hit an ASCII character.
44 return state == 0 ? VALID_ENDPOINT
45 : state == internal::I18N_UTF8_VALIDATOR_INVALID_INDEX
50 void StreamingUtf8Validator::Reset() {
54 bool StreamingUtf8Validator::Validate(const std::string& string) {
55 return StreamingUtf8Validator().AddBytes(string.data(), string.size()) ==